v0.1.0

The first release of preowned: a local HTTP server that serves the models built into macOS through the OpenAI API. Point any OpenAI client at http://127.0.0.1:8027/v1, and requests run on your Mac.

Endpoints

Endpoint	Model ID	Apple framework
GET /v1/models, GET /v1/models/{id}
POST /v1/chat/completions	apple-foundation-model	Foundation Models
POST /v1/images/generations	apple-image-playground	Image Playground
POST /v1/images/edits	apple-image-playground	Image Playground
POST /v1/images/variations	apple-image-playground	Image Playground
POST /v1/embeddings	apple-nl-embedding	NaturalLanguage
POST /v1/audio/transcriptions	apple-speech-transcriber	Speech
POST /v1/audio/translations	apple-speech-transcriber	Speech, then Translation
POST /v1/audio/speech	apple-speech-synthesizer	AVFoundation (AVSpeechSynthesizer)

OpenAI model names such as gpt-4o and whisper-1 map to the Apple model for each endpoint. The README lists what isn't supported, and Docs/compatibility.md covers every OpenAI parameter.

Requirements

• A Mac with Apple silicon and macOS 26 or later.
• Apple Intelligence turned on in System Settings, for chat and images.

Run preowned inventory to see which models work on your Mac and why the rest don't.

Install

brew install mattt/tap/preowned

The attached archive has the arm64 binary, so Homebrew doesn't need Xcode. To build from source instead, you need Xcode 27 or later; see the README.